1. Field of the Invention
The present invention relates to transmitting apparatuses and methods, receiving apparatuses and methods, programs and recording media used therewith, and transmitting/receiving systems, and in particular, to a transmitting apparatus and method, and a receiving apparatus and method, a program and a recording medium (which are used therewith), and a transmitting/receiving system in which high quality data can be obtained by using, for example, data (information) stored from the past.
In addition, the present invention relates to data processing systems, encoding apparatuses and methods, decoding apparatuses and methods, and programs and recording media used therewith, and in particular, to a data processing system, an encoding apparatus and method, a decoding apparatus and method, a program, and a recording medium which are used for reducing the amount of encoded data obtained by performing vector quantization on data.
2. Description of the Related Art
With the recent development in information communication technology, for example, regarding even personal digital assistances (PDAs) which are used when being carried, a type having a communication function has become common.
In many cases, PDAs perform wireless communication since they are portable. The data rate of the wireless communication is not such high speed as, for example, approximately 64 kilobits per second (bps). Accordingly, when one PDA transmits image data to another PDA, the data is transmitted, with its amount reduced, that is, deteriorated image data is transmitted.
Thus, it is difficult for the receiving PDA to obtain an image having high quality.
Also, in wireless communication, lack of partial transmission data, such as packets, easily occurs, and in this case, in the receiving PDA, the quality of the image more deteriorates.
By way of example, when data is encoded by vector quantization, a codebook in which code vectors are correlated with codes representing the code vectors is used to detect a code vector which minimizes the distance to the data, and the code is output as a vector-quantized result. The difference between the code vector represented by the code as the vector-quantized result and the original data is calculated to find a differential vector. The code as the vector-quantized and the differential vector are output as encoded data.
The codebook for use in vector quantization is created by performing learning based on the Linde Buzo Gray (LBG) algorithm (or the like) using, for example, a large amount of data for learning.
When data is encoded by vector quantization, the encoded data consists of a code and a differential vector. Thus, by reducing the differential vector as much as possible, the amount of the encoded data can be reduced.
However, the codebook has been used in fixed form, the differential vector is large depending on data. Thus, in some cases, it is difficult to reduce the amount of the encoded data.